Computerized assessment of syntactic complexity in Alzheimer’s disease: a case study of Iris Murdoch’s writing

Increase in Linguistic Complexity in Older Adults During COVID-19

The reported psychological impact of the COVID-19 pandemic and related public health measures included a decline in cognitive functioning in older adults. Cognitive functioning is known to correlate with the lexical and syntactic complexity of an individual's linguistic productions. We examined written narratives from the CoSoWELL corpus (v 1.0), collected from over 1,000 U.S. and Canadian older adults (55+ y.o.) before and during the first year of the pandemic. We expected a decrease in the linguistic complexity of the narratives, given the oft-reported reduction in cognitive functioning associated with COVID-19. Contrary to this expectation, all measures of linguistic complexity showed a steady increase from the pre-pandemic level throughout the first year of the global lockdown. We discuss possible reasons for this boost in light of existing theories of cognition and offer a speculative link between the finding and reports of increased creativity during the pandemic.

A curious case of retrogenesis in language: Automated analysis of language patterns observed in dementia patients and young children

Introduction

While linguistic retrogenesis has been extensively investigated in the neuroscientific and behavioral literature, there has been little work on retrogenesis using computerized approaches to language analysis.

Methods

We bridge this gap by introducing a method based on comparing output of a pre-trained neural language model (NLM) with an artificially degraded version of itself to examine the transcripts of speech produced by seniors with and without dementia and healthy children during spontaneous language tasks. We compare a range of linguistic characteristics including language model perplexity, syntactic complexity, lexical frequency and part-of-speech use across these groups.

Results

Our results indicate that healthy seniors and children older than 8 years share similar linguistic characteristics, as do dementia patients and children who are younger than 8 years.

Discussion

Our study aligns with the growing evidence that language deterioration in dementia mirrors language acquisition in development using computational linguistic methods based on NLMs. This insight underscores the importance of further research to refine its application in guiding developmentally appropriate patient care, particularly in early stages.

Automated Measures of Syntactic Complexity in Natural Speech Production: Older and Younger Adults as a Case Study

PURPOSE

Multiple methods have been suggested for quantifying syntactic complexity in speech. We compared eight automated syntactic complexity metrics to determine which best captured verified syntactic differences between old and young adults.

METHOD

We used natural speech samples produced in a picture description task by younger (n = 76, ages 18-22 years) and older (n = 36, ages 53-89 years) healthy participants, manually transcribed and segmented into sentences. We manually verified that older participants produced fewer complex structures. We developed a metric of syntactic complexity using automatically extracted syntactic structures as features in a multidimensional metric. We compared our metric to seven other metrics: Yngve score, Frazier score, Frazier-Roark score, developmental level, syntactic frequency, mean dependency distance, and sentence length. We examined the success of each metric in identifying the age group using logistic regression models. We repeated the analysis with automatic transcription and segmentation using an automatic speech recognition (ASR) system.

RESULTS

Our multidimensional metric was successful in predicting age group (area under the curve [AUC] = 0.87), and it performed better than the other metrics. High AUCs were also achieved by the Yngve score (0.84) and sentence length (0.84). However, in a fully automated pipeline with ASR, the performance of these two metrics dropped (to 0.73 and 0.46, respectively), while the performance of the multidimensional metric remained relatively high (0.81).

CONCLUSIONS

Syntactic complexity in spontaneous speech can be quantified by directly assessing syntactic structures and considering them in a multivariable manner. It can be derived automatically, saving considerable time and effort compared to manually analyzing large-scale corpora, while maintaining high face validity and robustness.

SUPPLEMENTAL MATERIAL

https://doi.org/10.23641/asha.24964179.

Using personal writings to detect dementia: A text mining approach

A novel text mining pilot for dementia detection using Linguistic Inquiry and Word Count (LIWC) was tested on public figures' writings looking at word choice and affect compared to those with and without dementia. The differences found in this analysis mirror the expected patterns where writings of people with dementia reflect significantly more analytical thinking words, but significantly less authentic and emotional tone. In addition, the analysis found that people with dementia use significantly less functional words, such as grammar, and affections (happiness, sadness, anger, sadness), but tend to use significantly more pronouns in their writings. Written samples of those with dementia also use significantly less time-oriented words that indicate past, present, or future. The analysis of free form text suggests a potential avenue for detecting early changes that correlate with dementia, allowing for early preventative treatment before noticeable cognitive impairment occurs.

Cognitive Writing Process Characteristics in Alzheimer's Disease

In this article, we explore if the observation of writing behavior can assist in the screening and follow-up of mild cognitive impairment (MCI) and mild dementia due to Alzheimer's disease (AD). To this end, we examined the extent to which overall writing process measures and pausing behavior during writing differed between 15 cognitively impaired patients and 15 age- and gender-matched healthy controls. Participants completed two typed picture description tasks that were registered with Inputlog, a keystroke logging program that captures keyboard activity during text production. The following variables were analyzed with mixed-effects models: time on task; number of characters, pauses and Pause-bursts per minute; proportion of pause time; duration of Pause-bursts; and pause time between words. For pause time between words, also the effect of pauses preceding specific word categories was analyzed. Results showed a main effect of group on all variables. In addition, for pause time between words a main effect of part-of-speech was found as well. Results indicate that writing process analysis can possibly serve as a supplementary tool for the screening and follow-up of AD.

A syntax-lexicon trade-off in language production

Spoken language production involves selecting and assembling words and syntactic structures to convey one's message. Here we probe this process by analyzing natural language productions of individuals with primary progressive aphasia (PPA) and healthy individuals. Based on prior neuropsychological observations, we hypothesize that patients who have difficulty producing complex syntax might choose semantically richer words to make their meaning clear, whereas patients with lexicosemantic deficits may choose more complex syntax. To evaluate this hypothesis, we first introduce a frequency-based method for characterizing the syntactic complexity of naturally produced utterances. We then show that lexical and syntactic complexity, as measured by their frequencies, are negatively correlated in a large (n = 79) PPA population. We then show that this syntax-lexicon trade-off is also present in the utterances of healthy speakers (n = 99) taking part in a picture description task, suggesting that it may be a general property of the process by which humans turn thoughts into speech.

Family history of FXTAS is associated with age-related cognitive-linguistic decline among mothers with the FMR1 premutation

BACKGROUND

Women who carry a premutation allele of the FMR1 gene are at increased vulnerability to an array of age-related symptoms and disorders, including age-related decline in select cognitive skills. However, the risk factors for age-related decline are poorly understood, including the potential role of family history and genetic factors. In other forms of pathological aging, early decline in syntactic complexity is observed and predicts the later onset of neurodegenerative disease. To shed light on the earliest signs of degeneration, the present study characterized longitudinal changes in the syntactic complexity of women with the FMR1 premutation across midlife, and associations with family history of fragile X-associated tremor/ataxia syndrome (FXTAS) and CGG repeat length.

METHODS

Forty-five women with the FMR1 premutation aged 35-64 years at study entry participated in 1-5 longitudinal assessments spaced approximately a year apart (130 observations total). All participants were mothers of children with confirmed fragile X syndrome. Language samples were analyzed for syntactic complexity and participants provided information on family history of FXTAS. CGG repeat length was determined via molecular genetic testing.

RESULTS

Hierarchical linear models indicated that women who reported a family history of FXTAS exhibited faster age-related decline in syntactic complexity than those without a family history, with that difference emerging as the women reached their mid-50 s. CGG repeat length was not a significant predictor of age-related change.

CONCLUSIONS

Results suggest that women with the FMR1 premutation who have a family history of FXTAS may be at increased risk for neurodegenerative disease, as indicated by age-related loss of syntactic complexity. Thus, family history of FXTAS may represent a personalized risk factor for age-related disease. Follow-up study is needed to determine whether syntactic decline is an early indicator of FXTAS specifically, as opposed to being a more general age-related cognitive decline associated with the FMR1 premutation.

Analysis and Classification of Word Co-Occurrence Networks From Alzheimer’s Patients and Controls

In this paper we construct word co-occurrence networks from transcript data of controls and patients with potential Alzheimer’s disease using the ADReSS challenge dataset of spontaneous speech. We examine measures of the structure of these networks for significant differences, finding that networks from Alzheimer’s patients have a lower heterogeneity and centralization, but a higher edge density. We then use these measures, a network embedding method and some measures from the word frequency distribution to classify the transcripts into control or Alzheimer’s, and to estimate the cognitive test score of a participant based on the transcript. We find it is possible to distinguish between the AD and control networks on structure alone, achieving 66.7% accuracy on the test set, and to predict cognitive scores with a root mean squared error of 5.675. Using the network measures is more successful than using the network embedding method. However, if the networks are shuffled we find relatively few of the measures are different, indicating that word frequency drives many of the network properties. This observation is borne out by the classification experiments, where word frequency measures perform similarly to the network measures.

Acceptability of collecting speech samples from the elderly via the telephone

OBJECTIVE

There is a critical need to develop rapid, inexpensive and easily accessible screening tools for mild cognitive impairment (MCI) and Alzheimer's disease (AD). We report on the efficacy of collecting speech via the telephone to subsequently develop sensitive metrics that may be used as potential biomarkers by leveraging natural language processing methods.

METHODS

Ninety-one older individuals who were cognitively unimpaired or diagnosed with MCI or AD participated from home in an audio-recorded telephone interview, which included a standard cognitive screening tool, and the collection of speech samples. In this paper we address six questions of interest: (1) Will elderly people agree to participate in a recorded telephone interview? (2) Will they complete it? (3) Will they judge it an acceptable approach? (4) Will the speech that is collected over the telephone be of a good quality? (5) Will the speech be intelligible to human raters? (6) Will transcriptions produced by automated speech recognition accurately reflect the speech produced?

RESULTS

Participants readily agreed to participate in the telephone interview, completed it in its entirety, and rated the approach as acceptable. Good quality speech was produced for further analyses to be applied, and almost all recorded words were intelligible for human transcription. Not surprisingly, human transcription outperformed off the shelf automated speech recognition software, but further investigation into automated speech recognition shows promise for its usability in future work.

CONCLUSION

Our findings demonstrate that collecting speech samples from elderly individuals via the telephone is well tolerated, practical, and inexpensive, and produces good quality data for uses such as natural language processing.

How to do things with (thousands of) words: Computational approaches to discourse analysis in Alzheimer's disease

Natural Language Processing (NLP) is an ever-growing field of computational science that aims to model natural human language. Combined with advances in machine learning, which learns patterns in data, it offers practical capabilities including automated language analysis. These approaches have garnered interest from clinical researchers seeking to understand the breakdown of language due to pathological changes in the brain, offering fast, replicable and objective methods. The study of Alzheimer's disease (AD), and preclinical Mild Cognitive Impairment (MCI), suggests that changes in discourse (connected speech or writing) may be key to early detection of disease. There is currently no disease-modifying treatment for AD, the leading cause of dementia in people over the age of 65, but detection of those at risk of developing the disease could help with the identification and testing of medications which can take effect before the underlying pathology has irreversibly spread. We outline important components of natural language, as well as NLP tools and approaches with which they can be extracted, analysed and used for disease identification and risk prediction. We review literature using these tools to model discourse across the spectrum of AD, including the contribution of machine learning approaches and Automatic Speech Recognition (ASR). We conclude that NLP and machine learning techniques are starting to greatly enhance research in the field, with measurable and quantifiable language components showing promise for early detection of disease, but there remain research and practical challenges for clinical implementation of these approaches. Challenges discussed include the availability of large and diverse datasets, ethics of data collection and sharing, diagnostic specificity and clinical acceptability.

Artificial Intelligence, Speech, and Language Processing Approaches to Monitoring Alzheimer's Disease: A Systematic Review

BACKGROUND

Language is a valuable source of clinical information in Alzheimer's disease, as it declines concurrently with neurodegeneration. Consequently, speech and language data have been extensively studied in connection with its diagnosis.

OBJECTIVE

Firstly, to summarize the existing findings on the use of artificial intelligence, speech, and language processing to predict cognitive decline in the context of Alzheimer's disease. Secondly, to detail current research procedures, highlight their limitations, and suggest strategies to address them.

METHODS

Systematic review of original research between 2000 and 2019, registered in PROSPERO (reference CRD42018116606). An interdisciplinary search covered six databases on engineering (ACM and IEEE), psychology (PsycINFO), medicine (PubMed and Embase), and Web of Science. Bibliographies of relevant papers were screened until December 2019.

RESULTS

From 3,654 search results, 51 articles were selected against the eligibility criteria. Four tables summarize their findings: study details (aim, population, interventions, comparisons, methods, and outcomes), data details (size, type, modalities, annotation, balance, availability, and language of study), methodology (pre-processing, feature generation, machine learning, evaluation, and results), and clinical applicability (research implications, clinical potential, risk of bias, and strengths/limitations).

CONCLUSION

Promising results are reported across nearly all 51 studies, but very few have been implemented in clinical research or practice. The main limitations of the field are poor standardization, limited comparability of results, and a degree of disconnect between study aims and clinical applications. Active attempts to close these gaps will support translation of future research into clinical practice.

Connected Speech Features from Picture Description in Alzheimer's Disease: A Systematic Review

The language changes that occur over the course of Alzheimer's disease (AD) can impact communication abilities and have profound functional consequences. Picture description tasks can be used to approximate everyday communication abilities of AD patients. As various methods and variables have been studied over the years, current knowledge about the most affected features of AD discourse in the context of picture descriptions is difficult to summarize. This systematic review aims to provide researchers with an overview of the most common areas of impairment in AD discourse as they appear in picture description tasks. Based on the 44 articles fulfilling inclusion criteria, our findings reflect a multidimensional pattern of changes in the production (speech rate), syntactic (length of utterance), lexical (word-frequency and use of pronouns), fluency (repetitions and word-finding difficulties), semantic (information units), and discourse (efficiency) domains. We discuss our findings in the light of current research and point to potential scientific and clinical uses of picture description tasks in the context of AD.

Genre-typical narrative arcs in films are less appealing to lay audiences and professional film critics

People tend to like stimuli-ranging from human faces to text-that are prototypical, and thus easily processed. However, recent research has suggested that less typical stimuli may be preferred in creative contexts, such as fine art or music lyrics. In an archival sample of movie scripts, we tested whether genre-typicality predicted film ratings as a function of rater role (novice audience member or expert film critic). Genre-typicality was operationalized as the profile correlations between linguistic arcs (across five segments, or acts) for each script and within-genre averages. We predicted (1) that critics would prefer more disfluent (genre-atypical) films and general audiences would prefer fluent (genre-typical) films, and (2) that these differences would be most pronounced for genres expected to be more entertaining (e.g., action/adventure) than challenging (e.g., tragedy). Partly consistent with our hypotheses, the results showed that critics gave higher ratings to action/adventure films with less typical positive emotion arcs. However, regardless of audience-member or professional-critic status, higher ratings were attributed to films that were more genre-atypical (or disfluent), in terms of analytic thinking, narrative action, and emotional tone, across all genres except family/kids films. Such findings support the growing literature on the appeal of disfluency in the arts and have relevance for researchers in psychology and computer science who are interested in computational linguistic approaches to attitudes, film, and literature.

Personal Journal Keeping and Linguistic Complexity Predict Late-Life Dementia Risk: The Cache County Journal Pilot Study

OBJECTIVES

We determined the feasibility of accessing personal journals and correlating markers of linguistic complexity with all-cause dementia and Alzheimer's disease (AD).

METHOD

A stratified random sample of 215 older adults reported on lifetime journal writing habits. From 66 of these participants (49% of those with journals), digital photographs of journal text were transcribed then subjected to the Linguistic Inquiry Word Count program to measure linguistic complexity markers: Words per Sentence, Percentage of 6+ Letter Words, Cognitive Mechanics, Percentage of Unique Words, and Percentage of Words that are Numerals. AD diagnosis was made via in-depth clinical protocol.

RESULTS

In the larger sample, ever being a journal writer significantly predicted a 53% reduction in all-cause dementia risk. In the subsample with transcribed writings, Percentage of 6+ Letter Words predicted AD and all-cause dementia risk, with all logistic regression models controlling for age, education, gender, and Latter-Day Saints affiliation.

DISCUSSION

These data suggest the potential viability of adulthood language use as a predictive tool for late-life AD risk, both in the linguistic features and the practice of journal writing itself.

Predicting probable Alzheimer's disease using linguistic deficits and biomarkers

BACKGROUND

The manual diagnosis of neurodegenerative disorders such as Alzheimer's disease (AD) and related Dementias has been a challenge. Currently, these disorders are diagnosed using specific clinical diagnostic criteria and neuropsychological examinations. The use of several Machine Learning algorithms to build automated diagnostic models using low-level linguistic features resulting from verbal utterances could aid diagnosis of patients with probable AD from a large population. For this purpose, we developed different Machine Learning models on the DementiaBank language transcript clinical dataset, consisting of 99 patients with probable AD and 99 healthy controls.

RESULTS

Our models learned several syntactic, lexical, and n-gram linguistic biomarkers to distinguish the probable AD group from the healthy group. In contrast to the healthy group, we found that the probable AD patients had significantly less usage of syntactic components and significantly higher usage of lexical components in their language. Also, we observed a significant difference in the use of n-grams as the healthy group were able to identify and make sense of more objects in their n-grams than the probable AD group. As such, our best diagnostic model significantly distinguished the probable AD group from the healthy elderly group with a better Area Under the Receiving Operating Characteristics Curve (AUC) using the Support Vector Machines (SVM).

CONCLUSIONS

Experimental and statistical evaluations suggest that using ML algorithms for learning linguistic biomarkers from the verbal utterances of elderly individuals could help the clinical diagnosis of probable AD. We emphasise that the best ML model for predicting the disease group combines significant syntactic, lexical and top n-gram features. However, there is a need to train the diagnostic models on larger datasets, which could lead to a better AUC and clinical diagnosis of probable AD.

Automated Assessment of Patients' Self-Narratives for Posttraumatic Stress Disorder Screening Using Natural Language Processing and Text Mining

Patients' narratives about traumatic experiences and symptoms are useful in clinical screening and diagnostic procedures. In this study, we presented an automated assessment system to screen patients for posttraumatic stress disorder via a natural language processing and text-mining approach. Four machine-learning algorithms-including decision tree, naive Bayes, support vector machine, and an alternative classification approach called the product score model-were used in combination with n-gram representation models to identify patterns between verbal features in self-narratives and psychiatric diagnoses. With our sample, the product score model with unigrams attained the highest prediction accuracy when compared with practitioners' diagnoses. The addition of multigrams contributed most to balancing the metrics of sensitivity and specificity. This article also demonstrates that text mining is a promising approach for analyzing patients' self-expression behavior, thus helping clinicians identify potential patients from an early stage.

Neural correlates of spelling difficulties in Alzheimer's disease

Alzheimer's disease (AD) is associated with a general cognitive decline that affects the memory and language domains. Thus, an oral production deficit with a lexical-semantic origin has been widely observed in these patients. Their written production capacities, however, have been much less studied. We assessed the spelling abilities of 22 AD patients and a group of matched healthy controls with a test battery including written picture naming and word and pseudoword dictation tests, as well as text dictation and spontaneous writing tasks. The results of the AD patients in the discriminative tasks were then entered into voxel-based morphometry analyses along with their grey matter volumes. The patient group presented a selective impairment for word dictation, which contrasted with a spared capacity to spell pseudowords, and showed more difficulties for words with arbitrary and rule-based orthography. Moreover, they also produced less complete syntactic units in the spontaneous writing task. These results point out the lexical-semantic, as opposed to sublexical, nature of the spelling deficit associated to AD. In addition, we recognized a mainly left-lateralized cortical network, including areas in the posterior inferior temporal lobe and the superior region of the parietal cortex, which might be responsible for this impairment. Other regions, such as the putamen, were also associated to the deficit. The results of this study, hence, improve our understanding of the neuropsychological and neuroanatomical mechanisms that underlie the cognitive symptoms associated to AD.